1 S E R I E manuales 32 Socio-economic impacts of natural disasters: a gender analysis Sarah Bradshaw Sustainable Development and Human Settlements División Women and Development Unit COOPERACION ITALIANA Santiago, Chile, May 2004
2 This document was prepared by Sarah Bradshaw, consultant for the Economic Commission for Latin America and the Caribbean (ECLAC), under the supervision of the Women and Development Unit, in close collaboration with the Sustainable Development and Human Settlements Division and the ECLAC Subregional Headquarters in Mexico City, in the framework of the Project Improve damage assessment methodology to promote natural disaster mitigation and risk reduction awareness and preparedness in Latin America and the Caribbean (ITA/99/130). This document has been reproduced without formal editing. The views expressed herein are those of the author and do not necessarily reflect the views of the Organization. United Nations Publications ISSN printed version: X ISSN electronic version: ISBN: LC/L.2128-P Sales number: E.04.II.G.56 Copyright United Nations, May All rights reserved. Printed in United Nations, Santiago, Chile Applications to the right to reproduce this work are welcome and should be sent to the Secretary of the Publication Board, United Nations Headquarters, New York, N. Y , USA. Member States and their Governmental Institutions may reproduce this work without prior authorization, but are requested to mention the source and inform the United Nations of such reproduction.
3 Index Abstract... 5 Introduction... 7 I. Methodology and approach... 9 A. Sources of information... 9 B. Basic concepts...10 II. Impacts of hurricane Mitch from a gender perspective...25 A. Direct impacts...25 B. Indirect impacts...30 III. Strategies for dealing with crisis situations and their consequences...35 A. Coping strategies...35 B. Reconstruction initiatives...37 IV. Summary and recommendations...45 A. Summary of the situation post hurricane Mitch...45 B. Recommendations...46 C. On information relating to vulnerable groups...47 D. On monitoring the impact of the disaster and subsequent interventions, from a gender perspective...49 E. On women s participation in the planning, design and monitoring of emergency programmes and rehabilitation projects...52 Bibliography...55 Serie Manuales: issues published
4 Socio-economic impacts of natural disasters: a gender analysis Index of tables Table 1 Types of vulnerability Table 2 Proposed vulnerability index using a gender approach Table 3 Gender-related development index Table 4 Aspects of a gender-focused analytical framework: holistic analysis of the situation facing people before and after disaster and response planning Table 5 Emotional impact and need for care in Nicaragua Table 6 Strategies for dealing with crises Table 7 Variables and relation in the process of change Index of figures Figure 1 Internal and external aspects of vulnerability Figure 2 Transformation and reconstruction: the theory Figure 3 Relations within households Figure 4 Productive work by women before and after hurricane Mitch in four nicaraguan communities by headship of households and by age (women in male-headed households) Figure 5 Change of roles-change of relations (simple model) Figure 6 Frustration and impact on conflict and violence
5 Abstract This paper analyses the socio-economic effects of hurricane Mitch using a gender approach and proposes new analysis indicators for crisis situations that may better reflect women s disadvantageous position relative to men. The first section of the document discusses key concepts used in gender and disaster analysis, in the context of the region and hurricane Mitch. The following section examines the direct and indirect impacts, and looks at how they have affected women, as well as the responses to Mitch at three levels: first, that of individuals and their strategies for coping with the crisis; second, the actions of governments and the coordinated bodies of civil society; and third, reconstruction initiatives carried out by national and international organizations. The final section attempts to draw together the salient points and challenges suggested by the analysis. It also offers some recommendations for integrating this approach into future emergency and reconstruction scenarios and for reducing women s current vulnerability. 5
6 Introduction Disasters tend to reveal existing national, regional and global power structures, as well as power relations within intimate relations (Enarson and Morrow, 1998:2). Hurricane Mitch emerged as a tropical depression on the afternoon of 21 October By the following morning it was already a grade 4 hurricane. Four days later, Mitch attained grade 5, becoming the fourth most intense hurricane in the Atlantic in the Twentieth Century. On 28 October, Mitch battered south-western Honduras, the Pacific coastline of Nicaragua and El Salvador. Mitch caused damage to Honduras 18 territorial departments. Strong winds wreaked havoc in the northern departments of Cortés and Colón. The lowlands in the east also experienced severe flooding, while flooding in the Sula Valley caused the Ulúa and Chamelecón Rivers to join, forming one river five miles wide that covered the cities of El Progreso, Tela and San Pedro Sula. The storm flooded the Choluteca Valley, as well as the cities of Comayagüela and Tegucigalpa. The heavy rains unleashed by Mitch battered the western and north-western parts of Nicaragua. Chinandega received an entire year s rainfall (1,600 mm) in the space of five days. The cities of Estelí, Madriz, Nueva Segovia and Matagalpa were flooded, and around 2,000 people died and 980 disappeared when part of the hillside of the Casitas volcano, located close to Chinandega, gave way, triggering a mudslide that reached speeds of up to 200 kilometres per hour and completely buried three small villages in the municipality of Posoltega. 7
7 Socio-economic impacts of natural disasters: a gender analysis As it moved northward, hurricane Mitch brought higher-than-usual rainfall, floods and high tides on the Salvadoran coast. The hurricane had a major impact on the Gulf of Fonseca, the Bajo Lempa Valley and the departments of La Unión and San Miguel, particularly in the Chilanguera River, which was where the majority of deaths in El Salvador occurred. It is estimated that the hurricane had a direct impact on one out of 10 people in the region and caused damage totalling US$ 4 billion to the productive sector, including two thirds of the infrastructure of Honduras and Nicaragua; US$ 1,2 billion in total physical infrastructure in the region; US$ 800 million in housing, health and education; and US$ 3 billion in raw materials and plantations (ECA 2000). Disasters such as Hurricane Mitch are natural phenomena, but their impacts are not. Rather, their effects are the result of the actions of human beings and are determined by the circumstances of the country in question, i.e., poverty, social inequalities, and the extent of deforestation, among other factors. Blaikie et al. (1994:3) point out that there is a risk inherent in treating disasters as something peculiar or as events detached from people s daily lives. In the countries of Central America, this serves as a warning against analyses that separate natural disasters from their political and socio-economic context, from economic growth in line with the neoliberal model, and from the vulnerabilities inherent in this process, which affect the impact of a hurricane such as Mitch. Moreover, where disasters take place in societies governed by power relations based on gender, age or social class, their impact will also reflect these relations and, as a result, people s experience of the disaster will vary. Today, the importance of the gender perspective during times of crisis and emergency is acknowledged, both on account of the differentiated impact on men and women and on account of the different strategies adopted by them to deal with such situations (Byrne, 1995; CAW, 1998). Despite the recent activities and publications of a small group of experts on disasters (Enarson and Morrow, 1998; Peacock et al., 1997), a gender perspective has not become mainstream within disasters research. Moreover, although a number of gender training manuals and guides have been made available, often the so called tyranny of the urgent wins out over good practice. 8
8 I. Methodology and approach A. Sources of information Up until early 2000, few studies had been conducted on the impact of Hurricane Mitch on the Central American region. The most important of these were three regional evaluations, with both a social and an economic focus, which constitute an important source of information: the first was commissioned by the Disasters Emergency Committee (DEC), of the United Kingdom (ECA, 2000), the second was issued by the Inter-American Development Bank (IDB) (Gomáriz, 1999) while the third was put out by the World Bank (WB) (Delaney and Shrader, 2000). In addition, a paper published by the Central American Symposium of Women involved in Reconstruction (CEM- H, 2000) provides an overview of the situation of women in the region, both before and after Mitch. This study is based on the analysis of existing documents and reports on the impact of Mitch, complemented by interviews with key informants. During the initial stages of this study, it became evident that there was a dearth of information available in general, and about the socio-economic situation of women in the countries of the region in particular. Two exceptions were however noted: The Survey on Gender, El Salvador (IUDOP, 1999). Carried out at the initiative of the Association of Women for Dignity and Life (Las Dignas), in coordination with other organizations, this survey reached some reliable conclusions about women s perceptions and opinions, and provided some 9
9 Socio-economic impacts of natural disasters: a gender analysis basic statistics. It included, inter alia, questions about violence, women s rights, women s attitudes as regards stereotypical roles, and information on environmental and vulnerability reduction initiatives in El Salvador. The Social Audit, Nicaragua (CCER, 1999a; 1999c). This survey, which was promoted by the Civil Coordinator for Emergency and Reconstruction, comprised two stages. The first stage, conducted in February 1999, covered all the departments affected by the hurricane and involved 10,000 interviews of households concentrated in the emergency period. The purpose of the questions was not only to estimate the damage sustained, but also to understand the perceptions of those affected by Mitch as regards the aid received (distribution, effectiveness, usefulness). The second stage, in September 1999, focused on the reconstruction process, and sought to quantify how much progress had been made, and learn about the actions of the various actors involved, and people s opinions about assistance. At the time this report was prepared, in September 2000, planning was under way on the third stage. To round out the data, the results of a primary investigation conducted in Nicaragua in 1999 were examined (Bradshaw et al., 2000); in addition, the study took into account the observation of activities carried out over a year and a half by participants in various institutions of organized civil society in Nicaragua. In the case of Honduras and El Salvador, the information is based on interviews of key respondents. Given that there are few experts on disasters with a gender focus, respondents were selected from among representatives of the women s and feminist movements who had some experience of Hurricane Mitch. This ensured there was an overall view of the situation, including a general analysis, examples of good and bad practices, discussion of the processes at the macro and national levels, and specific recommendations. B. Basic concepts Before the regional situation post Hurricane Mitch is examined, it is important to clarify a number of concepts. The issue of vulnerability in general, and that of the men and women of the countries in the region in particular, needs to be addressed. It is also essential to know what is meant by a gender perspective with respect to a focus on women. Lastly, it is also necessary to examine one of the most important places in times of crisis: the household. 1. Vulnerability Vulnerability is a key concept in predicting and understanding the existence of differentiated impacts on the various groups in a society (Blaikie et al., 1994). The concept of vulnerability takes into account people and the differences among them; in other words, it facilitates analysis of the social situation, affirming that people s circumstances change and can be changed, in this case, by an event such as a hurricane. Consequently, the concept does not look at the resources available to the different social groups in order to describe their current place in society (vulnerability as a passive concept), but rather to gauge the prospects for changing the situation (vulnerability as an active concept). The concept of vulnerability focuses on limitations or lack of access to resources; that notwithstanding, many investigations have attempted to foreground to a greater extent the positive, i.e., the uses to which people put the available resources, together with their self-help strategies in crisis situations. This notion of vulnerability considers the combination of external aspects the risks or intensity of an external shock or disaster (such as a hurricane) and internal aspects the ability to cope with such disasters without sustaining major damage, or ability to recover after the event. 10
10 Figure 1 INTERNAL AND EXTERNAL ASPECTS OF VULNERABILITY Vulnerability External aspects Intensity of the experience (depends on geographical location, etc.) Internal aspects Ability to recover after the disaster (depends on access to resources, etc.) a) Types of vulnerability Source: Author s elaboration. It is possible to identify different types of vulnerability: economic, social, political, physical, psychological. It is also necessary to take into account the intensity of the disaster and the ability to recover from it, though often the two are related. Table 1 sets out some different types of vulnerability and their definition. Situation involving lack of... Integration in the community (e.g., recently arrived migrant) Inclusion in the decision-making process (e.g., woman) Feeling of power or control over one s life (e.g., woman in a violent relationship) Feeling of security (e.g., high crime rate) Table 1 TYPES OF VULNERABILITY Leads to Social vulnerability Social vulnerability Psychological vulnerability Psychological vulnerability Health (e.g., Disabled person) Physical resources, such as money, housing, etc. Physical vulnerability Physical vulnerability Source: Author s elaboration. Vulnerability is directly related to impact. For example, Hurricane Mitch affected not only people who were deemed vulnerable, but also, as in the case of Tegucigalpa, the middle class. The loss of one s dwelling led to a situation of physical vulnerability (being homeless), social vulnerability (dependency on others) and psychological vulnerability (the trauma of the experience of loss), among others. However, even when these types of vulnerability are significant, especially for reconstruction plans, it is the vulnerabilities experienced by people prior to a disaster occurring that require the greatest attention. In this respect, case studies indicate that prior resources are reliable indicators of vulnerability, that is to say the extent of the damage caused by the impact and people s ability to bounce back (Enarson, 1998b). Research carried out by this author has identified major aspects of post-disaster survival and recovery: Income, savings, loans, insurance policies. Land, livestock, tools. 11
11 Socio-economic impacts of natural disasters: a gender analysis 12 Secure employment; work experience. Health and nutrition; food.security. Appropriate, secure housing. Functional education; administrative skills. Close family networks. Low rate of adult dependency in the household. Access to public and/or private transport. Time. Social networks; community integration. Political power and influence. Power in the household; access to, and control of, household resources. Access to emergency resources (information, shelters). These resources essential to survival and recovery are unequally distributed in all societies, which means that in equally dangerous environments, people and social groups are impacted in different ways. The following categories are the hardest hit (Enarson, 1998a): Poor and low-income households. Single-parent households. Socially isolated households. Recently arrived residents, immigrants, foreigners. Senior citizens, children and young people. People with a disease or a mental or physical disability. Undocumented residents; refugees; war veterans. Indigenous populations and subordinate ethnic groups. Institutionalized populations; homeless residents. Women. It is possible to predict based on Enarson s analysis that the following groups of women will be especially hard hit: Poor and low-income. Elderly. Having a disability or disease. Heads of household. Homeless. Indigenous. Immigrants. Isolated.
12 Rural. Affected by violence. Much of the work on gender and disasters, including the studies by Enarson, is based on research carried out in the United States. It is therefore important to evaluate to what extent this reflects the reality of the countries of Central America. With that aim in mind, it is useful to study a summary of the data drawn from various sources (Pan American Health Organization, PAHO; United Nations Children s Fund, UNICEF; International Organization for Migration, IOM) concerning female disaster victims and refugees in Honduras (Tábora, 2000). In regard to people made homeless by Mitch in Honduras, the data show the following: The majority are women (51%). The majority of the women are heads of household (51%), a figure which is significantly higher than the national percentage of female-headed households, according to the household survey (26%). A high proportion (37%) did not attended school or did not complete primary education; only 7% completed secondary education, and less than 1% had access to tertiary study. The women are mainly on low incomes. Children and teenagers account for more than half (56%) of all people taking refuge in shelters. 16% of women in shelters reported that they were pregnant. A significant number are heads of household and/or young single mothers, between 13 and 22 years of age (7% of all female-headed households in shelters). It is important to analyse this last point in light of the conclusions of a study conducted in Nicaragua (Bradshaw et al., 2000). This study found that, prior to Mitch, the situation of young women in independent households (female heads or young live-in partners) was worse than that of women aged over 25, in terms of their access to financial resources and perception of their own contribution, and that this situation worsened still further after Mitch, in both absolute and relative terms. This means that, in the context of the region, young women in independent households should be viewed a group in a particularly vulnerable situation. b) Vulnerability in the region A mapping of risks and vulnerability in the region, published shortly after the hurricane hit (Ordóñez et al., 1999), offers major insights into vulnerability in countries before Mitch. The conceptualization of vulnerability as seen by these authors involves three aspects: technical, political and social. The technical component is primarily concerned with infrastructure and the ability to resist the impact of the disaster. The second component, political vulnerability, is defined as the degree of a community s autonomy in respect of decision-making. It contends that the greater the autonomy, the lesser the extent of the community s political vulnerability. The third component is social vulnerability. The analysis is based on various elements used to measure the vulnerability indices of different countries in the region, along with various groups within them. These are: Poverty index. Health condition (mortality, morbidity). Malnutrition index. Percentage of households headed by women. Illiteracy index. 13
13 Socio-economic impacts of natural disasters: a gender analysis Housing conditions. These elements provide a solid foundation on which to analyse the situation in each country, but they do not provide for gauging the real situation of women in the region. One of the criticisms levelled at the study by Ordóñez et al. is that it uses a concept of vulnerability that is more passive than active, with a greater focus on the circumstances facing individuals than their prospects for recovery in the wake of a disaster. The problem is that it is more difficult to measure the opportunities or ability to use the resources available than it is to measure limitations or lack of resources. Placing the emphasis on abilities makes it necessary, at the least, to carry out studies in the local sphere, which means earmarking more resources for research in the region. However, the core of the problem is that the effectiveness of the research methods and the accuracy of the qualitative results have been called into question. For these reasons, the discussion that follows builds on the vulnerability components used by Ordóñez et al. and proposes the inclusion of some indicators that may provide a better picture of the situation facing women compared to that facing men. At the end of the section, a blueprint is outlined for measuring vulnerability from a gender perspective. (i) Poverty and access to, and control of, financial resources Poverty is important as a component of the vulnerability of individuals and the various social groups, but the way in which it is calculated limits its usefulness. In terms of gender, the methods that are generally used to measure poverty fail to reflect the unequal situation of women, especially within the home. Put differently, secondary poverty is as significant as poverty per se for women, because it reflects the fact, for example, that men do not hand over all their income to the household but rather use some of it for their social activities (drinking alcohol, among others). There may be cases of households that are not considered poor, in terms of income, but in which women and children actually live in (secondary) poverty because the resources available to them are far fewer than the total household resources. In this regard, a number of studies in the region indicate that it is very common for men to hold back income for themselves and that, on average, they allocate between 50% and 70% of their total income to the household (Chant, 1985, on Mexico; Bradshaw, 1995, on Honduras). The studies further point out that women use all or almost all of their income to satisfy the needs of the household and to care for their children. Therefore, women who do not have access to their own income constitute a vulnerable group. 1 In order to obtain data on secondary poverty, it is necessary to incorporate a component that reflects women s degree of financial dependence in households headed by a man. The proportion of male-headed households where the woman is not gainfully employed might be a proxy for secondary poverty. Lastly, where women are concerned, it is important to think more about access to, and control of, resources than about poverty per se. Again we run up against the difficulty of measuring control, hence the need to formulate a proxy. One possible solution would be to consider property ownership, i.e., the percentage of women who hold title to land or homes. In this way, it is also possible to reflect changes in how vulnerability is defined, when housing is identified as a potentially important resource for the generation of income and, by extension, for reducing vulnerability/poverty, as the World Bank has done in several of its studies. Moreover, the concept of poverty also reflects the way in which income is generated. Job security, together with the range of sources and options available to households to obtain jobs, are 1 Other studies show that the income generated by women living in male-headed households does not necessarily represent a supplement to men s contributions. Often men perceive such a situation as an opportunity for reducing still further their contributions, given that women now have money. 14
14 important in terms of vulnerability. A household that depends on a single source of income is more vulnerable than one that can count on a variety of sources for generating income. In addition, the source of employment is relevant when assessing the impact of a disaster. Accordingly, vulnerability mapping projects should include the most important sources of employment in each area, so as to ensure a rapid response in terms of economic recovery after any disaster. Lastly, there is a need to take into account the importance of credit for survival. In that regard, the impact of a disaster is twofold: it signifies the loss of items purchased on credit, such as seeds, as well as the loss of productive capacity for paying the debt. The ECA (2000) report states that, peasants also found themselves heavily in debt due to the emergency and some of them sold the bulk of their crops at very cheap prices, in order to be able to pay some of the debts related to financing the crops that Mitch destroyed. (ii) Health conditions General mortality and morbidity reveal the extent of health vulnerability in the wake of a disaster. However, it might also be useful to consider a number of indicators focusing on specific groups, such as women of child-bearing age and the population aged under 6, in relation to cases of maternal mortality and diarrhea, respectively. On account of their health status, it is also possible to identify other vulnerable groups in disaster situations: Abused women. People living with AIDS, who experience greater physical vulnerability not only due to their disease, but also because of their social status or socia l exclusion. AIDS indices in the region, especially Honduras, underline the importance of taking such people into consideration. Street children. In order to prevent discrimination against these and other vulnerable groups and be able to meet their specific requirements, it is necessary to implement, among others initiatives, awarenessbuilding and training programmes for all people involved in emergency situations (firemen, civil defence and Red Cross personnel). (iii) Malnutrition Though other indicators may be of use in measuring the degree of food security enjoyed by a particular population, the malnutrition index broken down by sex and age is an important indicator in itself. (iv) Female-headed households It is important to take into account the proportion of female-headed households, but the way in which this is done is subject to debate. Some experts believe that its importance lies in it being a line of analysis rather than as an indicator of vulnerability per se. Including the proportion of female heads of household as an indicator implies that it is not necessary to draw distinctions between women in urban and rural areas, young and older women, or among indigenous, and black and mixed-blood women. Advocates of this approach argue that the reality of vulnerability is the same for all these women, and that their vulnerability is different and more pronounced than that of women who live with a partner. Although the vulnerability of women heads of household is different from that of female spouses in male-headed households, both groups are in fact vulnerable (Bradshaw, 1996). Placing the spotlight on women heads of household may detract attention from those who live with a partner, which is the reality of most women. 15
15 Socio-economic impacts of natural disasters: a gender analysis (v) Illiteracy The illiteracy index constitutes a significant factor when gauging the population s opportunities for access to information, services and resources. However, it is also important to consider the degree of community organization, since this enables the factor of political vulnerability to be incorporated. Participation would be the most significant indicator, but presents conceptual problems. Furthermore, registries of non-governmental and community organizations are increasingly common in various countries and could serve as an indicator of the level of organization, but these were still not complete at the time this report was prepared. The inclusion of an indicator on the level of community preparedness for a disaster or, at least, awareness of the existence or lack of emergency/evacuation planning, should be an important indicator in this context. (vi) Housing conditions In addition to housing, the state of local infrastructure in general is also important; this covers schools, health posts, and especially transport links. Damage to highways may impact seriously on men and women s ability to recover from a disaster. For example, in Nicaragua, many women who were employed in buying and selling clothing were unable to continue doing so owing to the destruction of transport routes. Table 2 below presents a set of indicators that could be used as a vulnerability index for countries and communities, men and women, adults and young people. Poverty index Variable Households that are economically dependent on a man Lack of stable income Access to resources with the potential to generate income Range of sources of income Loans Mortality, morbidity Diarrhea in children aged less than 6 years Maternal mortality Malnutrition indices Housing conditions Table 2 PROPOSED VULNERABILITY INDEX USING A GENDER APPROACH Poverty Indicator Dissaggregated by the sex of the person heading household. Proportion of women without paid work in maleheaded households (proxy for secondary poverty). Proportion of the population that is unemployed or without stable employment, broken down by sex. Access to financial resources Proportion of the population with land and/or house title, broken down by sex (proxy for control of resources). Proportion of households with a single source of income, broken down by sex of the person employed. Proportion of households with loans, broken down by sex of person named. Health conditions Security Population broken down by sex. Population broken down by sex. Population broken down by age. Population broken down by sex. Population broken down by sex of the head of household. 16